A Theory of Constitution for Artifacts Abstract: I present a theory of constitution designed to apply to artifacts. It is based on the idea that artifacts are, in historical events, made out of their original constituting matter. This theory explains how constituted artifacts come into existence and I supply a theory of identity for artifacts based on the theory of constitution. Finally, I argue that the theories of constitution of Kit Fine and Lynne Baker fail to deal adequately with the constitution of artifacts, largely because they fail to incorporate the significance of the original events in which artifacts are made.
In this paper I will develop a theory of constitution as it applies to artifacts. By constitution I mean the relation that holds, paradigmatically, between a statue and what it is made out of. Whether there are examples of constitution outside the realm of artifacts, and if so, whether my theory holds in (or better: can be adapted to) such cases is something I will not address here. My suspicion is that the theory can be adapted to all those cases in which something like constitution obtains. In that case, my theory would be a general theory of constitution. But, if there are cases of constitution to which my theory cannot be adapted, I offer it merely as a theory of the relation of constitution as it obtains for artifacts. After I have expounded the theory, I will compare it with two other contemporary theories of constitution, those of Lynne Baker (2000, 2007) and Kit Fine (1982a). I will argue that to the extent that their theories diverge from mine, they face serious problems in analyzing the relation of constitution as it obtains in the case of artifacts.
I will make two important and related assumptions in this paper. One is that a three-dimensionalist account of material objects is correct. The second is that constitution is not identity: the statue and the copper are distinct, co-located objects. Both of these are controversial but I do not aim to defend them here. As I said, I will simply assume them. My goal is to understand artifactual constitution on the supposition that these assumptions are true. Since they are accepted by both Baker and Fine, the reasons I give for preferring my account to theirs do not in any way turn on these assumptions.
In my opening paragraph I described a paradigmatic case of constitution: a statue and the copper out of which it is made. The theory I shall develop is based on the idea of being made out of. I shall claim that:
C) x constitutes y if y is made out of x.
To understand C and to see why it may be plausible, we grasp what is involved in something’s being made out of something.1Being made out of is said in two ways. One of its senses is linked to an event in which a maker makes one thing out of another. It thus expresses a three-place relation involving a maker, an object out of which and an object into which (though we can elide mention of the maker). Unlike the other sense to be identified in a moment, this sense makes appropriate the use of continuous tenses and it allows both active and passive constructions. Standing in front of the sculptor at work we can say, “she is making a statue out of a piece of copper” or “a statue was being made out of a piece of copper when all hell let loose.” The other sense of the expression is a two-place relation between an object out of which and an object into which that is a possibly enduring state. It does not allow continuous tenses and active constructions. Standing in front of an origami model we can say, “this bird is made out of a single piece of paper.” We cannot, if the maker is not currently at work, say “this bird is being made out of a single piece of paper.” Where necessary, I shall distinguish the two senses by subscripting with an E (for the event use) or S (for the stative use). Constitution itself is a possibly enduring state between two objects. Therefore we will need the stative sense of being made out of to match constitution’s logical form. My suggestion, however, is that the stative sense is itself to be understood in terms of the event sense and that it is the properties of the event of making out of that lead to the characteristic features of the state of being made out of and ultimately of constitution.
There is another (possibly spurious) sense of being made out of that is akin to the stative sense. This is exhibited when we say of something that is made out of gold or made out of wood. Such cases I wish to exclude from consideration. For one thing, I am not sure that it isn’t even a solecism to say, of a copper statue, that is made out of copper. What we should say, I think, is that it is made of copper. And for some substances (not copper) we have special forms to denote the relation of being made of: “wooden,” “brazen,” “leaden,” “golden,” etc.2 A second reason for excluding this kind of being made [out] of concerns its relation to being madeE out of. To anticipate, I will argue, with numerous qualifications, that for something to be madeS out of something, it must have been madeE out of that thing. But if something is made [out] of copper this implication does not hold. Indeed, even things that are wholly unmade, like mountains, may be made [out] of granite or gold. Nor are such things as gold or copper the kinds of things a maker can direct her making activity towards. What she must direct her activity towards are such things as some copper or pieces of gold. What it means to say that an object was made [out] of copper is not that somebody once madeE it out of copper, but rather simply that at some point in the past, it would have been true to say of it “it is made [out] of copper.” Thus, as a matter of the pragmatics of speech acts, to say of an object that it was made [out] of copper implies that it is no longer made [out] of copper, whereas there is no such implication in saying that it was made out of a piece of copper or some copper. (Such a claim would naturally be taken as displaying the event sense of making out of.)
Having set this other use aside, I start the analysis, then, with the following:
M1) y is madeE out of x iff some agent A makesE y out of x.
A first stab at understanding what it is for an agent to makeE y out of x is:
A makesE y out x if and only if there is some kind F, such that y is an F, and x becomes y through A’s acting intentionally on x in order that it become an F.
This is essentially correct but one important refinement is needed. When A makesE y out of x, x has a certain kind of priority over y. In the typical case, this priority will be evident in the fact x pre-exists y. The agent does something to an already existing thing, x, which is not an F, in order that it may become an F. Even in non-divine cases, however, there are cases in which the priority of x over y must be sought elsewhere than in its pre-existence, since it does not pre-exist the object which is madeE out of it. For example, Allan Gibbard (1975, 94-6) talks about statue being made out of a piece of clay, which piece itself is created by the joining of two previous pieces in an act that puts them together and yields the statue all in one go. Analogously, Kit Fine (2003, 199) discusses a case of a statue’s being made out of some alloy, when the alloy is created by pouring two sources of molten metal into the cast: the statue and the alloy come into being at the same time. In such cases as these the priority of x over y is evident in the fact that the agent’s intentions are directed towards x and how making it a certain way will be sufficient, in the circumstances, for it to be an F. Since, in such cases, x does not become an F (since that would require its pre-existence), we should amend our definition thus:
M2) A makesE y out x if and only if there is some kind F, such that y is an F, and x is or becomes y through A’s acting intentionally on x in order that it be or become an F.3 At the heart of M2 is the idea that one thing can be or become another. In some sense, understanding this is the key to understanding constitution on my view. However, I shall delay further consideration of it until section II. In the discussion that now follows, I shall focus on the case of becoming rather than being just to simplify matters.
There are a couple of points to make about the aspect of intentionality in M2. Although the maker acts in such a way that x becomes y, she does not act with the intention of making x become y. This is because her intentions may exist prior to the act, and prior to the act there is no F that she intends to make x into. She acts in order to make x into an F, and in doing so, under the right circumstances, brings into existence y, a particular F. We should also note, and put to one side, uses of the expression “making out of” that do not require intention (or even require the absence of it), as when someone says “I made a fool out of myself,” or “you really made a mess out of that meal.” Such uses are idioms. There are also cases where the expression is meant literally but the act qualifies as making y out of x only by analogy with full-blown intentional cases. An example of this is when I absent-mindedly play with some putty and then give the result to my friend. Later I may say, “do you remember that thing I made out of some putty for you?”
We can now take a first step towards relating the two senses of making out of:
M3*) y is madeS out of x iff y is madeE out of x.
Filling this out from M1 and M2, this claims that y is madeS out of x if and only if for some F, and for some agent A, y is F and x was or became y though A’s acting intentionally on it in order that it be or become an F. This will require several refinements. The first refinement is that we must note an ambiguity in claims of being madeE out of. Often when we claim that someone makes y out of x, x is something that continues to exist. When, for example, a sculptor makesE a statue out of a piece of copper, the piece of copper continues to exist in addition to the statue. Sometimes, however, someone can makeE y out of x in such a way that x does not survive the event of the making of y. In this sense, a sculptor might make a statue out of a copper kettle by melting down the kettle and using the copper that it was madeS out of.4 If we include cases of this second sort, then M3* will fail in the right-to-left direction since the statue will have been madeE out a kettle but will not be madeS out of a kettle. Fortunately, in all cases in which y is madeE out of x and x does not survive, x itself is madeS out of something w, and it will also be true that y is madeE out of w. In makingE the statue out of the copper kettle, by melting it down, the sculptor is makingE the statue out of the copper that the kettle was madeS out of in such a way that the copper survives the making. So, if we restrict being madeE out of to include only cases in which the object out of which survives, we will not thereby fail to cover any objects that are madeS out of something.
It may seem as if, to understand how to take the right-hand side of M3*, we have invoked the notion being characterized on its left-hand side, being madeS out of, and hence rendered our account of makingS out of circular. This is not in fact so. We invoked makingS out of to re-assure ourselves that by understanding the right-hand side of M3* in the appropriately restricted way, we would not thereby fail to capture any cases which, intuitively, should fall under the analysis. Although the kettle is melted down in the making of the sculpture, we will not be precluded from saying that the statue is madeS out of something because in such cases, there is always something else out of which the made objects was madeE that does survive. Nonetheless, we do weaken M3*, or the theory of which it is a part, insofar as we now require, for its truth, the axiom:
A1) Wherever y is madeE out of x and x does not survive, there is another object w such that y is madeE out of w and w does survive.
Someone might challenge A1 by asking with what right we assume that when the statue is madeE out of the kettle (by melting it down) there is a piece of copper such that the statue is also madeE out of the piece of copper. A natural answer would be at hand if the kettle itself had been madeE out of a piece of copper. In that case, the piece of copper could have survived the making, the kettle itself would be madeS out of the piece of copper, and when it was melted down to make the statue, the copper out of which it was both madeS and madeE would survive and be used to makeE the new statue out of. But if the kettle itself had been made out of a statue that had been melted down, once again, with what right could we assume that it was madeE out of, and hence madeS out of, a piece of copper? So, suppose an infinitely extending backward chain of events in which a statue was made out of a kettle by melting it down, and the kettle was previously made out of a statue by melting it down, and that statue was previously made out of a kettle by melting it down, etc. It might be suggested that none of the items would have been madeE out of a piece of copper and hence, by M3*, none would be madeS out of a piece of copper. So we cannot assume, when the kettle is melted down, that there is a piece of copper that the statue is madeE out of. Thus we should accept neither A1 nor M3* which is supported by A1.
The possibility of such an infinite regress would certainly give grounds for doubting A1 though it does not imply its falsity since even with a regress, it might be true that when I melt down the kettle to make the statue, there is, notwithstanding the regress, a piece of copper that I make the statue out of. Nonetheless, A1 would certainly be better supported if we accepted as a further axiom the following:
A2) Everything that is madeE (and hence by M3*, madeS) out of something originates in something that is not madeE out of anything.
In other words, the creativity of makers must ultimately start with natural materials. A2 is a grounding axiom for makingE out of. It is, of course, implied by the assumption that there have not always been makers. (Actually, it is implied by something weaker: that no individual maker is part of an infinite series of causally linked makers.) Current science assures us that this is so and hence supports A2. But A2 is also independently plausible. I therefore adopt A1 and A2.5 For those who don’t want their metaphysics held hostage to science, however, or who do not find A2 independently plausible enough, we must acknowledge that without A1, M3* is somewhat uncertain; and that without A2, A1 is somewhat uncertain. Nothing, however, has shown that M3* or A1 is false under conditions of an infinite regress of making.
Returning to our discussion of M3*, a second necessary refinement is this. It is in the nature of such things as statues that they are able to change their matter over time. That means that such an object may come to be madeS out of something that it was not madeE out of. This challenges the left-to-right direction of M3*. To account for this, I characterize a notion of replacement. I shall look at some varieties of an informal notion of replacement and then characterize a technical sense of replacement in terms of them. One kind of replacement is easy to capture and lends its name to the generic notion of replacement. A piece of an artifact is either lost or broken and someone, possibly the original maker but also possibly another who acts as a restorer, replaces the lost or broken part with something suitable to sustain the existence and perhaps the use of the original object. Thus a restorer of a sculpture may replace the lost nose; the restorer of a car may replace a failed battery. In these cases, it is clear why the object continues to be made out of the new matter (which of course includes what remains unreplaced of the old matter), even though it was not originally made out of it. The process of replacement mimics in some respect the original making: it involves an agent who supplies the matter to do what the original matter did. Thus, whatever it is that allows the original maker to bring into existence a new object through her activity (and we have not yet looked at what that is) will allow the restorer to maintain that object in existence. Let us call this kind of replacement restoration. Another case is replacement by loss. The nose of the statue falls off; the statue is now made out of some copper that is not the original copper that it was made out of by the artisan. However, it continues to be made out of the new copper because the new copper is part of the original copper out of which it was originally made. Call this kind of replacement loss. A last case to consider is replacement by accretion. Accretion may be accidental or deliberate. A case of accidental accretion would be, for example, if a statue that is made out of dust, say, gets dusty. There is now a quantity of coherent dust slightly bigger than the original from which the statue was madeE. (Of course, copper statues get dusty too and since there is no intrinsic reason why a statue could not be madeS out of copper and dust, the problem arises there too.) In this case, the made object is not, I think, correctly described as being madeS out of the enlarged object; it continues to be madeS out of the original object to which something has been accidentally added. Deliberate accretion is another matter. (Think, for example, of adding a wing to an existing building.) Deliberate accretion seems similar to restoration in that it involves an agent acting intentionally in the light of the original object. There is, therefore, no objection to holding that the object comes to be madeS out of the enlarged object. So, let us define a technical notion of replacement:
x is a replacement of w =df x is either a replacement by loss of w, a replacement by restoration of w, or a replacement by deliberate accretion to w.
Not any amount of replacement guarantees that the remaining matter will be such that the original, or indeed any, object is madeS out of it. How much replacement a given object can endure may vary from kind to kind of constituted object. But it is important to realize that I am not here trying to give the conditions of continuity of a made object. Rather, I want to understand, given the existence of y and x, under what conditions y is madeS out of x. If y exists, and if x is the replacement by loss or restoration of that out of which y was madeE, then (subject to some other conditions), y is madeS out of x. The possibility of replacement by restoration or by loss raises another wrinkle, the other side of the coin of replacement. When original matter is replaced, it may continue to exist. We thus have a situation in which not only is the object now madeS out of something out of which it was not madeE, but in which it is not madeS out of that out of which it was originally madeE. In the extreme case, where all of the original matter has been replaced, we get the possibility of the Ship of Theseus puzzle. (A ship gradually has all its planks removed. These are then re-assembled into a ship resembling the original ship.) The ship is no longer madeS out the matter out of which it was originally madeE, but another ship is madeS out of that matter. (This description presupposes that I take the original ship to be the one with the new matter. I shall argue for this below.) We must thus stipulate that once an object’s original matter has been replaced, it is no longer madeS out of it.
We can now draw this material concerning replacement together to produce an improved version of M3*, to which we can also now add a temporal parameter:
M3) y is madeS out of x at t if and only if y is madeE out of x at t’ ( t) and x has not been replaced, or y is madeE out of w at t’and x is related by the ancestral of replacement to w.6 This takes care of both of the problems we noted above. In a case where the original matter has been replaced but still exists, as the original boards of the Ship of Theseus still exist after having all been replaced, the object does not count as (still) being madeS out of the original matter since it has been replaced. It does count as being madeS out of the new matter, although the agent didn’t makeE it out of that new matter, just in case the new matter is a replacement of the original matter (or the replacement of a replacement, etc.).
Having explained what is meant by being madeS out of, I can now state my primary thesis:
C) x constitutes y at t if y is madeS out of x at t.
This, as noted above, merely gives a sufficient condition for constitution. The question of whether the condition is also necessary, and if not, whether we can still claim to have captured the core of the concept of constitution, is something I shall not attempt to answer here.
Two things remain for me to complete the presentation of my artifactual theory of constitution. First, I have not yet addressed the crucial question of becoming: how, exactly, can one object become another without disappearing the in process? Secondly, what I have said so far says little about the conditions of identity of constituted objects and their essences. I shall take these up in the following sections.
In M2 I employed a notion of one object’s being or becoming another object. What does this mean and how is it possible? Philosophers (at least those sympathetic to constitution) have long recognized that the verb “to be” can, besides its other uses, be used to express a relation of constitution or something similar. “The statue is copper” means that the statue is made of copper, not that it is identical to copper or has the property of being copper. (This is the relation similar to constitution. I have, of course, excluded this relation from consideration in the current context.) “This piece of copper is a statue” means that the piece of copper constitutes a statue. If we add to the “is” of constitution a “becomes” of constitution, we seem to have the notions of being and becoming at work in M2. But of course, putting things this way suggests a circularity in the definition of constitution, since it is defined in terms of being madeS out of, which in turn is defined in terms of being madeE out of, which, it would now turn out, is defined in terms of the “is” and “becomes” of constitution. The mere fact that we could identity the “is” and “becomes” in M2 with the “is” and “becomes” of constitution does not make the account of constitution circular or vacuous. That would occur only if we could say no more about what we meant by those verbs than that they expressed the relation of constitution. We must, in other words, say more about them to keep our account of constitution from becoming vacuous. And indeed, we can. What our work up to now has accomplished is in pinpointing exactly where that something else must be supplied, and exactly what role that extra will play in an account of constitution.
To understand the “is” and “becomes” of MW, the question that really needs to be answered is how and why a new object comes into existence through the maker’s work on another object (already existing, in the case of becoming, to which I shall continue to confine myself for the sake of simplicity). And the answer to this question must steer a course between two extremes. On the one hand, it must show that something is at work that is powerful enough actually to bring into existence a new object on the back of the existing object. On the other hand, it must not invoke an ontological force the operation of which would commit us to many more objects than those we want for the account of makingE out of and constitution. Thus, for example, Peter van Inwagen (1990, sections 3-4) argues that there are no objects made out of other objects (at least partly) because the only non-arbitrary answers to when a new object can be brought into existence from other objects would imply a multitude of very implausible objects. And setting aside skepticism about the existence of anything other than ontological simples in general, there seem to be special reasons for worry in the case of artifacts about whether we can show how or why a new object comes into existence in making out of. Van Inwagen’s leading example of an artifact is a fort in the desert made by banking sand into walls. This striking image is presumably meant to underscore what he takes to be going on in the case of the making of all artifacts: that at bottom, such making can consist in nothing more than pushing already existing stuff around into various shapes.